Re: A brief history of (and agenda for) time zone libraries

Improvidently, a couple of things make buying back the GNU changes at this end difficult. One is not so bad. Since the GNU folks are clearly implementers (in the ANSI sense of the word), they're entitled to use names from the implementers' name space in their code; this means that strictly internal functions can be split into separate source files without polluting the user's name space. It's not clear that this is kosher when the timezone package is distributed independently. But some care in packaging and some creative use of the preprocessor would avoid most challenges here. One is not so tractable. Buying back the GNU code at this point would also involve putting things under the GPL, which I'm loathe to do. --ado